Optimization of the hölder image descriptor using a genetic algorithm

Local image features can provide the basis for robust and invariant recognition of objects and scenes. Therefore, compact and distinctive representations of local shape and appearance has become invaluable in modern computer vision. In this work, we study a local descriptor based on the Hölder exponent, a measure of signal regularity. The proposal is to find an optimal number of dimensions for the descriptor using a genetic algorithm (GA). To guide the GA search, fitness is computed based on the performance of the descriptor when applied to standard region matching problems. This criterion is quantified using the F-Measure, derived from recall and precision analysis. Results show that it is possible to reduce the size of the canonical Hölder descriptor without degrading the quality of its performance. In fact, the best descriptor found through the GA search is nearly 70% smaller and achieves similar performance on standard tests.

[1]  Cordelia Schmid,et al.  Local Grayvalue Invariants for Image Retrieval , 1997, IEEE Trans. Pattern Anal. Mach. Intell..

[2]  J. Lévy-Véhel.  Introduction to the multifractal analysis of images , 1998 .

[3]  Francisco Fernández de Vega,et al.  Selecting Local Region Descriptors with a Genetic Algorithm for Real-World Place Recognition , 2008, EvoWorkshops.

[4]  Leonardo Trujillo,et al.  Visual learning of texture descriptors for facial expression recognition in thermal imagery , 2007, Comput. Vis. Image Underst..

[5]  Jacques Lévy Véhel,et al.  Introduction to the Multifractal Analysis of Images , 1998 .

[6]  Pierrick Legrand,et al.  Local regularity-based interpolation , 2003 .

[7]  Leonardo Trujillo,et al.  Regularity based descriptor computed from local image oscillations. , 2007, Optics express.

[8]  Cordelia Schmid,et al.  A Performance Evaluation of Local Descriptors , 2005, IEEE Trans. Pattern Anal. Mach. Intell..

[9]  Cordelia Schmid,et al.  Evaluation of Interest Point Detectors , 2000, International Journal of Computer Vision.

[10]  Yan Ke,et al.  PCA-SIFT: a more distinctive representation for local image descriptors , 2004, CVPR 2004.

[11]  Stphane Mallat,et al.  A Wavelet Tour of Signal Processing, Third Edition: The Sparse Way , 2008 .

[12]  C. Tricot.  Curves and Fractal Dimension , 1994 .

[13]  Vincent Lepetit,et al.  Keypoint Signatures for Fast Learning and Recognition , 2008, ECCV.

[14]  Zehang Sun,et al.  Object detection using feature subset selection , 2004, Pattern Recognit..

[15]  Jack Sklansky,et al.  A note on genetic algorithms for large-scale feature selection , 1989, Pattern Recognit. Lett..

[16]  Pierrick Legrand,et al.  Local regularity-based image denoising , 2003, Proceedings 2003 International Conference on Image Processing (Cat. No.03CH37429).

[17]  S. Mallat.  A wavelet tour of signal processing , 1998 .

[18]  Leonardo Trujillo,et al.  Automated Design of Image Operators that Detect Interest Points , 2008, Evolutionary Computation.

[19]  Luc Van Gool,et al.  Speeded-Up Robust Features (SURF) , 2008, Comput. Vis. Image Underst..

[20]  Leonardo Trujillo,et al.  Scale Invariance for Evolved Interest Operators , 2009, EvoWorkshops.

[21]  Pierrick Legrand,et al.  Débruitage et interpolation par analyse de la régularité Hölderienne. Application à la modélisation du frottement pneumatique-chaussée , 2004 .

[22]  Jerzy W. Bala,et al.  Using Learning to Facilitate the Evolution of Features for Recognizing Visual Concepts , 1996, Evolutionary Computation.

[23]  Gustavo Olague,et al.  Evolutionary learning of local descriptor operators for object recognition , 2009, GECCO.

[24]  David G. Lowe,et al.  Object recognition from local scale-invariant features , 1999, Proceedings of the Seventh IEEE International Conference on Computer Vision.

[25]  Leonardo Trujillo,et al.  Synthesis of interest point detectors through genetic programming , 2006, GECCO.

[26]  S. Jaffard.  Wavelet Techniques in Multifractal Analysis , 2004 .

[27]  Francisco Fernández de Vega,et al.  Multiobjective design of operators that detect points of interest in images , 2008, GECCO '08.